I. Expressed Sequence Tag Nucleic Acid Molecules
Expressed sequence tags, or ESTs, are short sequences of randomly selected clones from a cDNA (or complementary DNA) library which are representative of the cDNA inserts of these randomly selected clones. McCombie, et al., Nature Genetics, 1: 124-130 (1992); Kurata, et al., Nature Genetics, 8: 365-372 (1994); Okubo, et al., Nature Genetics, 2: 173-179 (1992), all of which references are incorporated herein in their entirety.
Using conventional methodologies, cDNA libraries can be constructed from the mRNA (messenger RNA) of a given tissue or organism using poly dT primers and reverse transcriptase (Efstratiadis, et al., Cell 7:279-288 (1976), the entirety of which is herein incorporated by reference; Higuchi, et al., Proc. Natl. Acad. Sci. (U.S.A.) 73:3146-3150 (1976), the entirety of which is herein incorporated by reference; Maniatis, et al., Cell 8:163 (1976) the entirety of which is herein incorporated by reference; Land, et al., Nucleic Acids Res. 9:2251-2266 (1981), the entirety of which is herein incorporated by reference; Okayama, et al., Mol. Cell. Biol. 2:161-170 (1982), the entirety of which is herein incorporated by reference; Gubler, et al., Gene 25:263 (1983), the entirety of which is herein incorporated by reference).
Several methods may be employed to obtain full-length cDNA constructs. For example, terminal transferase can be used to add homopolymeric tails of dC residues to the free 3′ hydroxyl groups (Land, et al., Nucleic Acids Res. 9:2251-2266 (1981), the entirety of which is herein incorporated by reference). This tail can then be hybridized by a poly dG oligo which can act as a primer for the synthesis of full length second strand cDNA. Okayama and Berg, report a method for obtaining full length cDNA constructs. This method has been simplified by using synthetic primer-adapters that have both homopolymeric tails for priming the synthesis of the first and second strands and restriction sites for cloning into plasmids (Coleclough, et al., Gene 34:305-314 (1985), the entirety of which is herein incorporated by reference) and bacteriophage vectors (Krawinkel, et al., Nucleic Acids Res. 14:1913 (1986), the entirety of which is herein incorporated by reference; and Han, et al., Nucleic Acids Res. 15:6304 (1987), the entirety of which is herein incorporated by reference).
These strategies have been coupled with additional strategies for isolating rare mRNA populations. For example, a typical mammalian cell contains between 10,000 and 30,000 different mRNA sequences. Davidson, Gene Activity in Early Development, 2nd ed., Academic Press, New York (1976). The number of clones required to achieve a given probability that a low-abundance mRNA will be present in a cDNA library is N=(ln(1−P))/(ln(1−1/n)) where N is the number of clones required, P is the probability desired, and 1/n is the fractional proportion of the total mRNA that is represented by a single rare mRNA. (Sambrook, et al., Molecular Cloning: A Laboratory Manual, 2nd ed., Cold Spring Harbor Laboratory Press (1989), the entirety of which is herein incorporated by reference.).
A method to enrich preparations of mRNA for sequences of interest is to fractionate by size. One such method is to fractionate by electrophoresis through an agarose gel (Pennica, et al., Nature 301:214-221 (1983), the entirety of which is herein incorporated by reference). Another such method employs sucrose gradient centrifugation in the presence of an agent, such as methylmercuric hydroxide, that denatures secondary structure in RNA (Schweinfest, et al., Proc. Natl. Acad. Sci. (U.S.A.) 79:4997-5000 (1982), the entirety of which is herein incorporated by reference).
A frequently adopted method is to construct equalized or normalized cDNA libraries (Ko, Nucleic Acids Res. 18:5705-5711 (1990), the entirety of which is herein incorporated by reference; Patanjali, S. R. et al., Proc. Natl. Acad. Sci. (U.S.A.) 88:1943-1947 (1991), the entirety of which is herein incorporated by reference). Typically, the cDNA population is normalized by subtractive hybridization. Schmid, et al., J. Neurochem. 48:307-312 (1987) the entirety of which is herein incorporated by reference; Fargnoli, et al., Anal. Biochem. 187:364-373 (1990) the entirety of which is herein incorporated by reference; Travis, et al., Proc. Natl. Acad. Sci. (U.S.A.) 85:1696-1700 (1988) the entirety of which is herein incorporated by reference; Kato, Eur. J. Neurosci. 2:704 (1990); and Schweinfest, et al., Genet. Anal. Tech. Appl. 7:64 (1990), the entirety of which is herein incorporated by reference). Subtraction represents another method for reducing the population of certain sequences in the cDNA library. Swaroop, et al., Nucleic Acids Res. 19:1954 (1991), the entirety of which is herein incorporated by reference).
ESTs can be sequenced by a number of methods. Two basic methods may be used for DNA sequencing, the chain termination method of Sanger et al., Proc. Natl. Acad. Sci. (U.S.A.) 74: 5463-5467 (1977), the entirety of which is herein incorporated by reference and the chemical degradation method of Maxam and Gilbert, Proc. Nat. Acad. Sci. (U.S.A.) 74: 560-564 (1977), the entirety of which is herein incorporated by reference. Automation and advances in technology such as the replacement of radioisotopes with fluorescence-based sequencing have reduced the effort required to sequence DNA (Craxton, Methods, 2: 20-26 (1991), the entirety of which is herein incorporated by reference; Ju et al., Proc. Natl. Acad. Sci. (U.S.A.) 92: 4347-4351 (1995), the entirety of which is herein incorporated by reference; Tabor and Richardson, Proc. Natl. Acad. Sci. (U.S.A.) 92: 6339-6343 (1995), the entirety of which is herein incorporated by reference). Automated sequencers are available from, for example, Pharmacia Biotech, Inc., Piscataway, N.J. (Pharmacia ALF), LI-COR, Inc., Lincoln, Nebr. (LI-COR 4,000) and Millipore, Bedford, Mass. (Millipore BaseStation).
In addition, advances in capillary gel electrophoresis have also reduced the effort required to sequence DNA and such advances provide a rapid high resolution approach for sequencing DNA samples (Swerdlow and Gesteland, Nucleic Acids Res. 18:1415-1419 (1990); Smith, Nature 349:812-813 (1991); Luckey et al., Methods Enzymol. 218:154-172 (1993); Lu et al., J. Chromatog. A. 680:497-501 (1994); Carson et al., Anal. Chem. 65:3219-3226 (1993); Huang et al., Anal. Chem. 64:2149-2154 (1992); Kheterpal et al., Electrophoresis 17:1852-1859 (1996); Quesada and Zhang, Electrophoresis 17:1841-1851 (1996); Baba, Yakugaku Zasshi 117:265-281 (1997), all of which are herein incorporated by reference in their entirety).
ESTs longer than 150 bases have been found to be useful for similarity searches and mapping. (Adams, et al., Science 252:1651-1656 (1991), herein incorporated by reference.) EST sequences normally range from 150-450 bases. This is the length of sequence information that is routinely and reliably generated using single run sequence data. Typically, only single run sequence data is obtained from the cDNA library, Adams, et al., Science 252:1651-1656 (1991). Automated single run sequencing typically results in an approximately 2-3% error or base ambiguity rate. (Boguski, et al., Nature Genetics, 4:332-333 (1993), the entirety of which is herein incorporated by reference).
EST databases have been constructed or partially constructed from, for example, C. elegans (McCombrie, et al., Nature Genetics 1: 124-131 (1992), human liver cell line HepG2 (Okubo, et al., Nature Genetics 2:173-179 (1992)), human brain RNA (Adams, et al., Science 252:1651-1656 (1991); Adams, et al., Nature 355:632-635 (1992)), Arabidopsis, (Newman, et al., Plant Physiol. 106:1241-1255 (1994)); and rice (Kurata, et al., Nature Genetics 8:365-372 (1994).
II. Sequence Comparisons
A characteristic feature of a protein or DNA sequence is that it can be compared with other known protein or DNA sequences. Sequence comparisons can be undertaken by determining the similarity of the test or query sequence with sequences in publicly available or propriety databases (“similarity analysis”) or by searching for certain motifs (“intrinsic sequence analysis”)(e.g. cis elements)(Coulson, Trends in Biotechnology, 12: 76-80 (1994), the entirety of which is herein incorporated by reference; Birren, et al., Genome Analysis, 1: 543-559 (1997), the entirety of which is herein incorporated by reference).
Similarity analysis includes database search and alignment. Examples of public databases include the DNA Database of Japan (DDBJ) (available on the Worldwide Web at ddbj.nig.ac jp/); Genebank (available on the Worldwide Web at ncbi.nlm.nih.gov/web/Genbank/Index.htlm); and the European Molecular Biology Laboratory Nucleic Acid Sequence Database (EMBL) (available on the Worldwide Web at ebi.ac.uk/ebi_docs/embl_db.html). A number of different search algorithms have been developed, one example of which are the suite of programs referred to as BLAST programs. There are five implementations of BLAST, three designed for nucleotide sequences queries (BLASTN, BLASTX, and TBLASTX) and two designed for protein sequence queries (BLASTP and TBLASTN) (Coulson, Trends in Biotechnology, 12: 76-80 (1994); Birren, et al., Genome Analysis, 1: 543-559 (1997)).
BLASTN takes a nucleotide sequence (the query sequence) and its reverse complement and searches them against a nucleotide sequence database. BLASTN was designed for speed, not maximum sensitivity, and may not find distantly related coding sequences. BLASTX takes a nucleotide sequence, translates it in three forward reading frames and three reverse complement reading frames, and then compares the six translations against a protein sequence database. BLASTX is useful for sensitive analysis of preliminary (single-pass) sequence data and is tolerant of sequencing errors (Gish and States, Nature Genetics, 3: 266-272 (1993), the entirety of which is herein incorporated by reference). BLASTN and BLASTX may be used in concert for analyzing EST data (Coulson, Trends in Biotechnology, 12: 76-80 (1994); Birren, et al., Genome Analysis, 1: 543-559 (1997).
Given a coding nucleotide sequence and the protein it encodes, it is often preferable to use the protein as the query sequence to search a database because of the greatly increased sensitivity to detect more subtle relationships. This is due to the larger alphabet of proteins (20 amino acids) compared with the alphabet of nucleic acid sequences (4 bases), where it is far easier to obtain a match by chance. In addition, with nucleotide alignments, only a match (positive score) or a mismatch (negative score) is obtained, but with proteins, the presence of conservative amino acid substitutions can be taken into account. Here, a mismatch may yield a positive score if the non-identical residue has physical/chemical properties similar to the one it replaced. Various scoring matrices are used to supply the substitution scores of all possible amino acid pairs. A general purpose scoring system is the BLOSUM62 matrix (Henikoff and Henikoff, Proteins, 17: 49-61 (1993), the entirety of which is herein incorporated by reference), which is currently the default choice for BLAST programs. BLOSUM62 is tailored for alignments of moderately diverged sequences and thus may not yield the best results under all conditions. Altschul, J. Mol. Biol. 36: 290-300 (1993), the entirety of which is herein incorporated by reference, uses a combination of three matrices to cover all contingencies. This may improve sensitivity, but at the expense of slower searches. In practice, a single BLOSUM62 matrix is often used but others (PAM40 and PAM250) may be attempted when additional analysis is necessary. Low PAM matrices are directed at detecting very strong but localized sequence similarities, whereas high PAM matrices are directed at detecting long but weak alignments between very distantly related sequences.
Homologues in other organisms are available that can be used for comparative sequence analysis. Multiple alignments are performed to study similarities and differences in a group of related sequences. CLUSTAL W is a multiple sequence alignment package available that performs progressive multiple sequence alignments based on the method of Feng and Doolittle, J. Mol. Evol. 25: 351-360 (1987), the entirety of which is herein incorporated by reference. Each pair of sequences is aligned and the distance between each pair is calculated; from this distance matrix, a guide tree is calculated, and all of the sequences are progressively aligned based on this tree. A feature of the program is its sensitivity to the effect of gaps on the alignment; gap penalties are varied to encourage the insertion of gaps in probable loop regions instead of in the middle of structured regions. Users can specify gap penalties, choose between a number of scoring matrices, or supply their own scoring matrix for both the pairwise alignments and the multiple alignments. CLUSTAL W for UNIX and VMS systems is available at: ftp.ebi.ac.uk. Another program is MACAW (Schuler et al., Proteins, Struct. Func. Genet, 9:180-190 (1991), the entirety of which is herein incorporated by reference, for which both Macintosh and Microsoft Windows versions are available. MACAW uses a graphical interface, provides a choice of several alignment algorithms, and is available by anonymous ftp at: ncbi.nlm.nih.gov (directory/pub/macaw).
Sequence motifs are derived from multiple alignments and can be used to examine individual sequences or an entire database for subtle patterns. With motifs, it is sometimes possible to detect distant relationships that may not be demonstrable based on comparisons of primary sequences alone. Currently, the largest collection of sequence motifs in the world is PROSITE (Bairoch and Bucher, Nucleic Acid Research, 22: 3583-3589 (1994), the entirety of which is herein incorporated by reference.) PROSITE may be accessed via either the ExPASy server on the World Wide Web or anonymous ftp site. Many commercial sequence analysis packages also provide search programs that use PROSITE data.
A resource for searching protein motifs is the BLOCKS E-mail server developed by S. Henikoff, Trends Biochem Sci., 18:267-268 (1993), the entirety of which is herein incorporated by reference; Henikoff and Henikoff, Nucleic Acid Research, 19:6565-6572 (1991), the entirety of which is herein incorporated by reference; Henikoff and Henikoff, Proteins, 17: 49-61 (1993). BLOCKS searches a protein or nucleotide sequence against a database of protein motifs or “blocks.” Blocks are defined as short, ungapped multiple alignments that represent highly conserved protein patterns. The blocks themselves are derived from entries in PROSITE as well as other sources. Either a protein or nucleotide query can be submitted to the BLOCKS server; if a nucleotide sequence is submitted, the sequence is translated in all six reading frames and motifs are sought in these conceptual translations. Once the search is completed, the server will return a ranked list of significant matches, along with an alignment of the query sequence to the matched BLOCKS entries.
Conserved protein domains can be represented by two-dimensional matrices, which measure either the frequency or probability of the occurrences of each amino acid residue and deletions or insertions in each position of the domain. This type of model, when used to search against protein databases, is sensitive and usually yields more accurate results than simple motif searches. Two popular implementations of this approach are profile searches (such as GCG program ProfileSearch) and Hidden Markov Models (HMMs) (Krough et al., J. Mol. Biol. 235:1501-1531 (1994); Eddy, Current Opinion in Structural Biology 6:361-365 (1996), both of which are herein incorporated by reference in their entirety). In both cases, a large number of common protein domains have been converted into profiles, as present in the PROSITE library, or HHM models, as in the Pfam protein domain library (Sonnhammer et al., Proteins 28:405-420 (1997), the entirety of which is herein incorporated by reference). Pfam contains more than 500 HMM models for enzymes, transcription factors, signal transduction molecules, and structural proteins. Protein databases can be queried with these profiles or HMM models, which will identify proteins containing the domain of interest. For example, HMMSW or HMMFS, two programs in a public domain package called HMMER (Sonnhammer et al., Proteins 28:405-420 (1997)) can be used.
PROSITE and BLOCKS represent collected families of protein motifs. Thus, searching these databases entails submitting a single sequence to determine whether or not that sequence is similar to the members of an established family. Programs working in the opposite direction compare a collection of sequences with individual entries in the protein databases. An example of such a program is the Motif Search Tool, or MoST (Tatusov et al. Proc. Natl. Acad. Sci. 91: 12091-12095 (1994), the entirety of which is herein incorporated by reference.) On the basis of an aligned set of input sequences, a weight matrix is calculated by using one of four methods (selected by the user); a weight matrix is simply a representation, position by position in an alignment, of how likely a particular amino acid will appear. The calculated weight matrix is then used to search the databases. To increase sensitivity, newly found sequences are added to the original data set, the weight matrix is recalculated, and the search is performed again. This procedure continues until no new sequences are found.